Microbial Community Structure Drives Predatory Myxobacteria Distribution Under Different Compost Manures

Myxobacteria are unique predatory microorganisms with a distinct social lifestyle. The associated taxa play key roles in the microbial food webs in different ecosystems and regulate the community structures of soil microbial communities. Compared with conditions under conventional management, under organic conditions, myxobacteria abundance increases in the soil, which could be related to the presence of abundant myxobacteria in the applied compost manure. In the present study, high-throughput sequencing technologies were used to investigate the distribution patterns and drivers of predatory myxobacteria community distribution patterns in four common compost manures. According to the results, there was a signicant difference in predatory myxobacteria community structure among different compost manure treatments (P < 0.05). The alpha-diversity indices of myxobacteria community under swine manure compost were the lowest (Observed OTU richness = 13.25, Chao1 = 14.83, Shannon = 0.61), and those under wormcast were the highest (Observed OTU richness = 30.25, Chao1 = 31.65, Shannon = 2.62). Bacterial community diversity and Mg 2+ and Ca 2+ concentrations were the major factors inuencing myxobacteria distribution patterns under different compost manure treatments. In addition, pH, total nitrogen, and organic carbon inuenced myxobacteria distribution in compost manure. The predator–prey relationship between prey bacteria and myxobacteria and the interaction between myxobacteria and specic bacterial taxa (Micrococcales) in compost manure could explain the inuence of bacteria on myxobacteria community structure. Further investigations on the in-situ distribution patterns of predatory myxobacteria and the key bacteria inuencing their distribution are would advance our understanding of the ecological distribution patterns and functions of predatory microorganisms in the microbial world.


Introduction
Today, predation is considered a major evolutionary and ecological driver that can in uence community structure and ecosystem function [1][2][3]. Although extensive research has been carried out on predation behavior of large organisms such as animals and plants, predation is much less understood in the microbial world [3]. Myxobacteria are the rst taxa of bacteria described as micro-predators, which are capable of secreting antibiotics, hydrolases and bacteriolytic compounds to kill and lyse their prey microorganisms, including bacteria, fungi, protozoa, and other microorganisms [2,[4][5][6]. Myxobacteria are reportedly highly adaptable cosmopolitans distributed in diverse environments, such as nutrient-rich soil, compost, sand, rocky soil, freshwater lake mud, and the sea [7][8][9][10]. However, our understanding of the factors driving the distribution of predatory myxobacteria communities under different environmental conditions remains poor.
Interactions between myxobacteria and prey populations are a key aspect of the microbial food web [11], and predatory myxobacterium controls cucumber Fusarium wilt by regulating soil microbial community structure via decrease modularity and community number and increase connection number per node [12].
In addition, myxobacteria Corallococcus sp. strain EGB can prey on diverse soil bacteria, and could in uence microbial community structure in a microcosm system [13]. The accumulation of prey-speci c predacious genes in myxobacteria genomes partly explains the broad range of myxobacteria prey [14], which makes it reasonable to speculate that the bacterial community structure (composition and abundance) in microbial ecological niches could in uence the distribution patterns of myxobacteria communities.
In the Anthropocene, manure, which is mainly excreted by animals or derived from plant residues, is an environmentally friendly soil amendment used to manage soil degradation, and, in turn, increase crop production stability and agroecosystem functioning [15]. Manure application improves soil biochemical properties, such as soil organic carbon, total nitrogen (TN), and microbial biomass carbon [16]. Previous studies have demonstrated that myxobacteria naturally exist in compost manure [10,17,18], and that the diversity of soil microorganisms and myxobacteria increases under organic farming conditions [19,20]. In addition, environmental factors could in uence the ecological distribution of myxobacteria in soil [21]; however, it is unclear whether the application of compost manure rich in myxobacteria would increase myxobacteria abundance in the soil. In addition, whether environmental factors directly in uence the distribution of myxobacteria in compost manure, or indirectly in uence myxobacteria distribution via their effects on prey diversity in compost require further investigations.
The aim of the present study was to investigate the factors driving myxobacteria distribution patterns under four common compost manures. The speci c objectives were 1) to investigate the distribution patterns of predatory myxobacteria under different compost manures, 2) to explore the abiotic factors in uencing the distribution patterns of predatory myxobacteria in compost manure, 3) to explore the correlation between predatory myxobacteria diversity and bacterial community diversity under different compost manures, and 4) to explore the role of predatory myxobacteria in the bacterial community networks in compost manure, and the correlations with associated bacterial groups.

Sample collection and analysis
Four types of compost manure were used in the analyses in the present study. Cow dung (CD, Qingshen Zhongxing Farming Professional Cooperative in Meishan city, Sichuan Province, China), swine manure (SM, Tenghui Farming in Yingtan city, Jiangxi Province, China), and chicken manure (CM, Longxiang Farming in Suqian city, Jiangsu Province, China) samples were collected from commercial composts produced by aerobic fermentation. Wormcast manure (WC) obtained by vermicomposting wormcast for one month, was collected from our laboratory (produced by earthworms ingesting cow dung samples).
Each compost manure treatment included four replicates, and the different compost samples were divided into two parts, placed on ice, and transferred to the laboratory. One part was stored at -20℃ for use in DNA extraction and microbial analyses. The other part was stored at 4℃ for use in the determination of compost physicochemical properties. The pH, TN, organic matter, ammonium-nitrogen (NH 4 + -N), nitrate-nitrogen (NO 3 --N), total potassium (TK), and total phosphorus (TP) were determined based on methods described in a previous study [19].
Quali ed reads were processed using the Quantitative Insights Into Microbial Ecology pipeline [23]. TrimGalore (http://www.bioinformatics.babraham.ac.uk/projects/trim_galore/) and Flash [24] were used to process the nal V3-V4 tag sequences. Operational taxonomic units (OTUs) were clustered based on a 97% similarity cut off using UPARSE (version 7.1 http://drive5.com/uparse/) using a novel "greedy" algorithm that performs chimera-ltering and OTU-clustering simultaneously [24]. The taxonomy of each 16S rRNA gene sequence was analyzed using the RDP Classi er algorithm (http://rdp.cme.msu.edu/) against the Silva 16S rRNA database with a con dence threshold of 70%. The sequence numbers in each sample were normalized to the smallest sample size using the"normalized. shared"command in Mothur [25]. High-throughput sequencing data have been deposited in the National Center for Biotechnology Information Sequence Read Archive (BioProject ID PRJNA723459, study accession number SRP14294735-SRP14294746).

Predatory myxobacteria community abundance
Based on the 16S rRNA gene-based OTU results, a total of 682 OTUs annotated as Myxococcales (myxobacteria) at the order level were selected from all the 26412 bacterial OTUs clustered. Predatory myxobacteria community abundance information was obtained based on the abundance information for the 682 myxobacteria OTUs.

Statistical analysis
The α-diversity (Observed OTU richness, Chao1, Shannon, ACE, and Simpson) of myxobacteria and bacteria were estimated using Mothur [26]. The hierarchical cluster tree was calculated using the "vegan" package in R v 3.5.1(R Foundation for Statistical Computing, Vienna, Austria). Based on the abundance of the bacterial and myxobacteria OTUs, principal coordinate analysis (PCoA) and multivariate regression trees (MRT) were performed and constructed, respectively, using the "vegan" and "mvpart" packages, respectively, in R v 3.5.1 (R Foundation for Statistical Computing), Linear regression analysis was used to test the correlation between myxobacteria diversity and bacterial diversity, with sample variety as a random effect, myxobacteria diversity as the dependent variable and bacterial diversity as the independent variable. Analysis of the differences in the abundance of myxobacteria OTUs was performed using the "edgeR" package in R v3.5.1 [27]. Network analysis was performed using single/double factor correlation network analysis, and les were generated for network graph visualization using free online platforms, Majorbio I-Sanger Cloud Platform (http://www.i-sanger.com) and MicrobiomeAnalyst (https://www.microbiomeanalyst.ca).
Structural equation modeling (SEM) was used to explore the direct and indirect effects of physicochemical parameters and bacterial community diversity on myxobacteria community structure using the "lavaan" package in R v3.5.1 [28]. Our SEM analysis included six variables: myxobacteria diversity, bacterial diversity, pH, TN, Mg 2+ , and Ca 2+ . Distributions of myxococcales in the four types of manure compost were assessed using the "vcd" package in R v 3.5.1 and Circos (http://circos.ca/). Other statistical analyses were conducted using SPSS 13.0 (SPSS Inc., Chicago, IL, US).
Alpha-diversity indices (Observed OTU richness, Chao1, and Shannon diversity) values for bacterial and myxobacteria communities in different compost manures are illustrated in Figure 1. One-way Analysis of Variance results showed that the compost manures altered bacterial and myxobacteria community diversity and abundance signi cantly (P < 0.05) ( Figure 1, Table S1). The relative abundance and diversity of bacteria differed signi cantly among the four types of compost manure (P < 0.05) (Table S1).
In the case of myxobacteria communities, the alpha-diversity indices were the lowest in SM (Observed OTU richness = 13.25, Chao1 = 14.83, Shannon = 0.61), and the highest in WC (Observed OTU richness = 30.25, Chao1 = 31.65, Shannon = 2.62). Although myxobacteria abundance in CM was higher than that in CD, myxobacteria diversity exhibited opposite trends in the two compost manures ( Figure 1B). In addition, myxobacteria abundance and diversity trends in all four types of compost manure were similar to those of bacteria ( Figure 1A).

Myxobacteria Community Structure Among the Four Compost Manures
Based on the relative abundances of bacteria in different compost manures, the main bacterial orders in all samples were Micrococcales, Xanthomonadales, Clostridiales, Anaerolineales, Flavobacteriales, Rhizobiales, Sphingobacteriales, Pseudomonadales, and Myxococcales ( Figure S1). Myxococcales was a major taxa in the bacterial communities, accounting for approximately 3.15% of the total bacteria in the four types of compost manure.
At the family level, within the myxobacteria communities, the dominant families (merging small taxa with counts < 10, Figure 2A) across the four types of compost manure were Haliangiaceae (37%), OM27 (5%), Polyangiaceae (4%), and Nannocystaceae (2%). Haliangiaceae abundance in the CD and SM was higher than those in the other two compost manure types. Conversely, Polyangiaceae relative abundance was the highest in CD. At the order level, the relative abundance of Myxococcales was signi cantly different among the four types of compost manure (P < 0.05) ( Figure 2B, Table S2), while CD had the highest abundance (CD > WC > CM >SM, 74% > 19% > 6.1% > 0.78%).
According to the PCoA (carried out based on Bray-Curtis distances) plots ( Figure 3A) and Dendrogram Analysis (carried out based on Bray-Curtis distances) results ( Figure 3B), myxobacteria community structure was signi cantly different among the four types of compost manure (P < 0.05) (Table S2). In the rst component of the PCoA analysis (PCoA1), the community structures in CM and WC were rather similar, and the hierarchical clustering trees showed similar results ( Figure 3).

Correlations Between Myxobacteria and Bacterial Community Diversity
The community distribution of myxobacteria in compost manure was in uenced signi cantly by bacterial community diversity. There were signi cant and positive linear correlations between myxobacteria and bacterial community diversity (α-and β-diversity) in the four types of compost manure (Figure 4, P < 0.0001). The PCoA1 axes of myxobacteria abundance and bacteria abundance showed signi cant linear relationships with each other ( Figure 4A, R 2 = 0.9986, P < 0.0001), and, among the multiple diversity indices, there were consistent results with regard to Shannon diversity between myxobacteria and bacteria ( Figure 4B, R 2 = 0.94411, P < 0.0001).

Network analysis and structural equation modeling of myxobacteria community structure in compost manures
An ecological network illustrates the interaction of various organisms in an ecosystem. In the correlation network in the present study, the symbiotic relationship between myxobacteria and other bacteria drove the ecological network topology. As illustrated in Figure 6A, a single factor correlation network with 20 nodes was constructed based on the four types of compost manure under study (Table S5). In the network, myxobacteria and other bacteria (order level) formed a complex topological network structure (absolute value of Spearman's correlation coe cient ≥ 0.6). The Myxococcales node had a relatively high degree and clustering coe cient, and co-occurred with some nutrition-related bacteria; Myxococcales had a signi cant and positive correlation with bacterial orders (Cellvibrionales, Sphingomonadales, Flavobacteriales, Burkholderiales, Cytophagales, Rhodospirillales, and Rhizobiales), and a signi cant and negative correlation with Micrococcales ( Figure S3).
According to the results of two-factor correlation network analysis, the abundance of myxobacteria was

Discussion
Soil bacteria biogeography could reveal signi cant correlations between bacterial community distribution and environmental factors [29]. Moreover, microorganism distribution and development are in uenced by complex interactions with plants, animals, and other microbes, which could have bene cial, neutral, or harmful effects on bacterial community members [6]. In the present study, we investigated the distribution of predatory myxobacteria in microbial communities in different compost manures. Overall, the results indicated that bacterial community diversity, Mg 2+ and Ca 2+ concentrations, and pH were associated with myxobacteria community diversity in different compost manures.

Myxobacteria Community Structure in Different Compost Manures
Myxobacteria are mainly distributed in soil environments and most predatory myxobacteria are isolated from agricultural soils [2]. Compared with conventional farming, organic farming with organic fertilizer amendment can enhance microbial diversity and richness [19]. The application of organic fertilizer can signi cantly increase the diversity and richness of myxobacteria in the soil [20]. According to a previous study, myxobacteria in a single soil sample accounted for 4.1% of the entire bacterial community [21], and Myxococcale sequences accounted for 1.31−4.17% of the sequences in 16S rRNA gene libraries in the four types of compost manure examined in the present study. The relative abundance of myxobacteria in farmland soils [19] and subtropical and tropical forest soils [6] have been reported to account for 0.36-4.10% and 1.49-4.74% of the total bacterial abundance, respectively, which are consistent with the results reported in the present study under compost manure.
Myxobacteria were unevenly distributed in different compost manures, and not all myxobacteria families could be observed in the samples examined; particularly, Myxococcaceae and Cystobacteraeae were not observed in some samples. Cystobacterineae are frequently isolated from environments using culturedependent methods [30]; however, according to our results, Cystobacterineae abundance in the different compost manures was low, potentially highlighting the bias of the culture-independent method.
Researchers have reported that Sorangium can secrete high amounts of cellulose-degrading enzymes [31].
In the present study, Sorangium had high relative abundance in the CM compost, which could be associated with the high cellulose amount in the CM compost. In addition, in the present study, myxobacteria diversity in WC manure was signi cantly higher than that in SM. This could be because compared to SM, WC contains more easily usable organic substances [32], and higher microbial diversity and abundance [33], which provides adequate food and a suitable environment for myxobacteria development and survival.

Effects of Abiotic Factors on Myxobacteria Community Structure in Compost Manure
According to previous reports, microbial community structure is mainly in uenced by environmental factors [34], and soil characteristics are correlated with myxobacteria abundance [21]. Similarly, myxobacteria distribution in the different compost manures in the present study was in uenced by abiotic factors. We also observed correlations between abiotic factors of compost manure and myxobacteria community structure. In addition, according to the SEM results, abiotic factors of manure compost in uenced myxobacteria community structure directly.
According to the results of the present study, abiotic factors signi cantly in uence the distribution patterns of predatory myxobacteria in different compost manures. Speci cally, pH, TN concentration, and Mg 2+ concentration are signi cantly positively correlated with myxobacteria community diversity; conversely, Ca 2+ and NH 4 + -N concentrations in compost manure were signi cantly and negatively correlated with myxobacteria community diversity. pH is the major abiotic factor in uencing the distribution of microorganisms in different environments [35][36][37][38]. Similarly, pH considerably in uenced myxobacteria distribution in compost manure in the present study. Myxobacteria are mostly distributed in environments with a pH of approximately 6.5-8.5, especially in neutral to weakly alkaline soils with pH 6.0-8.0 [39,40]. Excluding in the case of CD (pH 9.21-9.47), the pH values of the other three compost manures (pH 7.09-7.82) were all within the optimal range for the myxobacteria survival, which could explain the high myxobacteria abundance in the three compost manures. NH 4 + -N is the preferred nitrogen source for most microorganisms [41]. Therefore, NH 4 + -N could positively in uence bacterial community structure when agricultural waste compost is adopted as fertilizer [34]; however, in the present study, we observed that myxobacteria abundance in compost manure was signi cantly negatively correlated with NH 4 + -N concentration, which is similar to results on the distribution of predatory bacteria in soil [20].
Salt ion concentrations can in uence the growth and development of myxobacteria [42][43][44][45]. Mg 2+ and Ca 2+ concentrations are generally considered to promote myxobacteria development [46,47]. Notably, in the present study, Mg 2+ concentrations in different compost manures were positively correlated with myxobacteria diversity, while opposite trends were observed with regard to Ca 2+ concentrations. Based on the SEM analysis results, Mg 2+ not only had a positive effect on myxobacteria diversity in different compost manures but also indirectly in uenced myxobacteria diversity by affecting bacterial diversity.
In the present study, high Ca 2+ concentrations were the major reason for the decrease in myxobacteria diversity in the four types of compost manure, with Ca 2+ concentration being negatively correlated with myxobacteria diversity; however, it can positively affect bacterial diversity in the four types of compost manure. In addition, the complex interactions between myxobacteria and indigenous microorganisms cannot be overlooked. High concentrations of Ca 2+ could promote the growth of some bacteria antagonistic to myxobacteria and, in turn, in uence myxobacteria diversity. According to our results, high Ca 2+ concentrations were signi cantly and positively correlated with Micrococcales abundance, while Micrococcales abundance was signi cantly and negatively correlated with Myxococcales abundance and diversity, which is partly consistent with our speculation.

Effect of Manure Compost bacterial diversity on Myxobacteria Community Structure
Microbial interactions such as auxotrophies and nutrient demands among members of a microbial community are key drivers of microbial community structure [48]. As a major class of predatory bacteria in microbial communities, myxobacteria can prey on gram-negative and gram-positive bacteria, yeasts, fungi, protozoa, and nematodes [2,49]. We observed signi cant correlation between myxobacteria diversity and bacterial community composition under different compost manures. Although no causal relationship was established, the results revealed direct correlation between potential prey and predators at the community level.
Because soil myxobacteria cannot synthesize ribo avin and branched-chain amino acids [3], their community structures could be in uenced by prey availability. In the present study, there was a positive correlation between the relative abundance of bacteria and the relative abundance of myxobacteria in compost manure. Considering predation of myxobacteria on prey bacteria, the reason for the increase in myxobacteria abundance is potentially an increase in the number of prey bacterial cells. Notably, some researchers have reported that Corallococcus sp. EGB strains control cucumber wilt disease by migrating to plant roots and regulating soil microbial community structure at the sites [12]. Consistent with our results, other studies on predator-prey diversity relations have reported positive correlation between predator abundance and prey abundance [50][51][52]. To the best of our knowledge, this is the rst study to explore the relationship between predatory myxobacteria and bacterial community structure across different types of compost manure.
Preferential predation by micro-predators could explain the in uence of bacterial communities on predatory microbial community structure. According to the results of the single-factor correlation network analysis in the present study, in the microbial correlation network of the different compost samples, Myxococcales were signi cantly positively correlated with Sphingomonadales, Flavobacteriales, Cellvibrionales, Cytophagales, Burkholderiales, Rhizobiales, and Rhodospirillales (order level), which implies potential predator-prey relationships. The gram characteristics of prey bacteria could in uence prey selection by myxobacteria. Myxococcus are reportedly more supported by gram-negative prey than by gram-positive bacteria [53]. We also note that in the present study, Myxococcales nodes in the correlation network were signi cantly and positively correlated with gram-negative bacteria, and signi cantly and negatively correlated with gram-positive bacteria (Micrococcales). However, the overall conclusion that gram-negative prey can more effectively support Myxococcales requires further investigation and evidence. Other studies have also reported that Haliangiaceae are effectively supported by Arthrobacter globiformis (a gram-positive actinomycete) [5]. In the present study, we highlight the in uence of bacterial community structure on myxobacteria community distribution patterns in different compost manures. Nevertheless, the key bacterial taxa driving the distribution patterns of myxobacteria require further investigations.

Conclusions
In the present study, we reveal the key factors in uencing myxobacteria distribution in different compost manures for the rst time. We report that abiotic factors (pH and Mg 2+ ) have positive effects on of myxobacteria community diversity as well as bacterial community diversity. However, high Ca 2+ concentrations have negative effects on myxobacteria diversity. Overall, bacterial community diversity and Mg 2+ and Ca 2+ were the major factors in uencing myxobacteria distribution in different compost manures. Nevertheless, due to the complex predator-prey interactions, our data failed to determine the speci c bacterial groups in uencing the myxobacteria distribution in the compost manures. Our ndings could facilitate the selection of appropriate compost manure types and appropriate management soil management strategies based on the physicochemical properties of the compost, which could not only increase myxobacteria community diversity and abundance in manure compost but also enhance soil health in farmland amended with organic fertilizer.
Declarations diversity on multitrophic interactions in a biodiversity experiment.